Skip to content

Benchmark

VBench 1.0 human action

Scores are percentages.

Models measured
9
Scores recorded
9
Sources
1
IQ weight
0.28

Score distribution

96.1%
Mean
96.2%
Median
2.6%
Std dev
92.8%
10th percentile
99.4%
90th percentile
92.4%
Min
92.4%99.6%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.28 for video-generation models, measured over 8 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Vidu Q12025-04-17· VBench data docs99.6%
2Veo 3· VBench data docs99.4%
3Sora· VBench data docs98.2%
4Vidu Q4· VBench data docs97.4%
5Kling V1.6· VBench data docs96.2%
6Mochi 1· VBench data docs94.6%
7Hunyuan Video2025-05-22· VBench data docs94.4%
8LTX Video5s 768×512· VBench data docs92.8%
9Video 01· VBench data docs92.4%

Data credits: VBench (Vchitect), MIT (licence)